LTX 2.3 Pro is one of only six models on Picasso IA confirmed to generate NSFW content directly, without a safety filter standing between your prompt and the render. Describe a fictional adult scene and get back a short cinematic clip, with camera motion and lighting that hold together from frame to frame. It works only with fictional, adult characters, on an adults-only platform.
Most video models bolt on a content filter that rejects anything it reads as adult, even when the scene is fictional and the account belongs to a verified adult. LTX 2.3 Pro is different. It is one of only six models on Picasso IA confirmed to generate NSFW content directly, without a safety filter standing between your prompt and the render. That confirmation comes straight from how the model is wired into the app, and it is the reason this page exists instead of pointing you toward a model that will just refuse the job halfway through.
What makes LTX 2.3 Pro worth reaching for beyond that is the footage itself. It renders motion the way a short film clip moves, with camera pans, depth, and lighting that hold together across frames instead of flickering or drifting. For adult fiction specifically, that consistency matters more than almost anywhere else, because a scene that warps between frames breaks the illusion the whole clip is trying to build. LTX 2.3 Pro was built for cinematic output first, and it happens to also be one of the few models here willing to render adult scenes without blocking the attempt.
If you only need a still image, the same confirmed, filter-free approach exists on the image side too, through models like Seedream 4.5, Qwen Image 2, Recraft V4.1 Pro, and Picasso Image Editor Pro. For video, your two confirmed options are LTX 2.3 Pro and Picasso Video. This page focuses on LTX 2.3 Pro because its camera work is the closest thing on the platform to a directed shot rather than a looping clip. Reach it any time from the toolkit, where every model on Picasso IA lives behind one shared interface.
LTX 2.3 Pro is a text to video model. You describe a scene in plain language, it renders a short clip that follows that description, and the result reads as a shot rather than a slideshow of images stitched together. Its strength is in how it treats a scene as a continuous space: a camera can push in, pan across a room, or hold on a subject while the lighting and shadows stay consistent from the first frame to the last. That is the main thing that separates a convincing clip from an obviously synthetic one.
Because it is one of the confirmed filter-free models, LTX 2.3 Pro will attempt an adult fictional prompt directly rather than bouncing it back to you. That does not mean every prompt produces a perfect result on the first try, video generation is still an iterative process, and the model can misread an ambiguous instruction the same way any generative model can. What it does mean is that a well-written adult scene involving fictional, adult characters gets rendered instead of refused outright, which is the specific gap this model fills on Picasso IA.
Duration and resolution are configurable per generation, and the exact options, along with what each combination costs in credits, are shown right in the generation panel before you commit. Check the live options in the toolkit when you start a project, since those settings determine your cost and render time. If you later want to touch up a specific frame as a still image, the same account also gives you access to Picasso Image Editor Pro for adult-safe still edits.
Getting from an idea to a finished clip takes a handful of steps, all inside the toolkit.
That is the entire loop. Most people run through several drafts of the same scene before landing on the one they keep.
A prompt for video needs more structure than one for a still image, because you are describing motion and time, not just a single frame. A few habits make a real difference in the output:
Describe the shot like a director would. Name the subject, the setting, and the action in that order, then add how the camera behaves, static, panning, pushing in, or orbiting. LTX 2.3 Pro responds well to that kind of layered instruction because it was built to render continuous camera motion, not just a single composed frame.
Anchor the lighting and mood early in the prompt. Words like "soft evening light" or "dim candlelight" give the model a consistent target to hold across the whole clip, which keeps a scene from looking assembled from mismatched frames.
Keep the scene physically plausible. A prompt that contradicts itself, two incompatible settings in one sentence, or an action that does not fit the space described, tends to produce a confused result. One clear scene per prompt beats several ideas crammed together.
Iterate in small steps. If a render is close but not right, change one detail at a time rather than rewriting the whole prompt, so you can tell which change fixed the issue and keep your credit spend efficient.
Reference your own uploads when it helps. Attaching a reference image for a character's appearance alongside your text prompt gives the model something concrete to match, which produces more consistent results across multiple generations of the same character.
Picasso IA is an adults-only platform, and every account generating adult content on it is expected to be an adult using it for fictional, adult material. That is a firm line, not a suggestion: content depicting a real person's likeness, a public figure, or anyone who could be read as a minor is never allowed, regardless of the model, the prompt wording, or the intent behind it. LTX 2.3 Pro being filter-free for adult fiction does not change that boundary, it simply means fictional adult scenes are not blocked the way they would be on a model that filters all NSFW content outright.
It is also worth understanding what happens on a model that is not one of the six confirmed filter-free options, models like Nano Banana, Nano Banana Pro, Seedream 5 Pro, Sora 2, Veo 3.1, or Kling. Those models apply their own content filters, and a prompt that reads as adult will often get refused. When that happens, Picasso IA shows a "Content Not Supported by This Model" message and recommends switching to one of the confirmed models instead, carrying your prompt over automatically. That is the intended path when a filtered model turns down your prompt: switch to a model built for it, such as LTX 2.3 Pro, rather than rephrase your way around the filter. Disguising a prompt to slip past a filter is not something this platform supports, and doing so does not change the rule about who and what can be depicted.
If you are building a longer piece with recurring characters, keep a clear internal description of each one and lean on reference images where you can, which helps both the quality of your results and your ability to keep every character clearly fictional and clearly adult.
Picasso IA runs on a credit system, and new accounts start with free credits, so you can generate your first clips with LTX 2.3 Pro before spending anything. Beyond that starting balance, credits are purchased as needed or included with a subscription plan, and the exact packages, tiers, and current free-credit amount are kept up to date on the pricing page rather than repeated here, since those numbers change over time.
Video generation cost is not a flat number. Because LTX 2.3 Pro lets you choose duration and resolution per render, the credit cost scales with what you select, a short, lower-resolution clip costs less than a longer, higher-resolution one. That cost is shown in the generation panel before you submit, updating live as you change the settings, so there is never a surprise charge after the fact.
If your budget is tighter for a project, render a shorter draft first to check the prompt and camera work, then commit to a longer, higher-resolution version once the scene works the way you want. That avoids paying for a long render that needed a prompt fix anyway. The same credit balance covers every other model on the platform, so credits spent exploring the full catalog are never locked to one model.
Yes. LTX 2.3 Pro is one of six models on Picasso IA confirmed to generate adult content directly, without a safety filter refusing the request, as long as the scene involves fictional, adult characters.
No separate plan is required. LTX 2.3 Pro is available to any account with enough credits to cover the render, whether from the free starting balance, a purchase, or a subscription. Current plan details are on the pricing page.
No. Picasso IA does not allow content depicting a real person's likeness or a public figure, on LTX 2.3 Pro or any other model, regardless of how the prompt is worded. Every adult generation on the platform must involve fictional characters.
Models that apply a content filter, which is most models on the platform, show a "Content Not Supported by This Model" message when a prompt reads as adult. Picasso IA then recommends switching to a confirmed filter-free model, including LTX 2.3 Pro, and carries your prompt over automatically.
No, it shares that status with Picasso Video, the only other confirmed filter-free video model on the platform. LTX 2.3 Pro is generally the stronger pick when camera movement and cinematic pacing matter most to the shot.
Render time depends on the duration and resolution you select, with longer, higher-resolution clips taking more time. You can watch the render progress directly in the toolkit.
Yes, attaching a reference image alongside your text prompt is supported and often improves consistency, especially for a character's appearance across multiple generations.
For fictional adult still images in the same project, the confirmed filter-free image models are Seedream 4.5, Qwen Image 2, Recraft V4.1 Pro, and Picasso Image Editor Pro, all reachable from the toolkit.
Cost scales with the duration and resolution you choose for the render, not with a flat per-model rate. A short, lower-resolution clip is cheaper than a long, high-resolution one, and the exact credit cost is shown before you generate.
You can regenerate with an adjusted prompt to change or extend a scene, and for still-frame touch-ups you can move into Picasso Image Editor Pro using a frame from your render as a starting point.
